What Isaac Newton Did in Lockdown – And What it Tells Us About Data Science

While sitting in the warm stillness of the orchard at Woolsthorpe Manor, you never would have guess that the surrounding towns and cities of England were gripped by illness and panic. The year was 1665, and Isaac Newton had been forced to abandon his studies at Cambridge, following a nationwide outbreak of plague. There he was, self-isolating at his parents’ home in Lincolnshire, when he observed an apple falling from a tree. It caused young Newton to ponder. Then it prompted him to experiment. Then it compelled him to write. And the end result was his theory of universal gravitation, set out in his landmark work: ‘Philosophiæ Naturalis Principia Mathematica’.
 
So the story goes -- though some details are debateable. For example, there’s no evidence that the fruit actually landed on Newton’s head. But one thing’s for sure – if England hadn’t been in the middle of a 17^th century lockdown, he may never have formulated his ideas about gravity. Or calculus and optics, for that matter.
 
History is crammed with similar stories. At times of extremity – like war, famine, plague, or indeed a coronavirus pandemic – new circumstances force people to think and act differently. And that can lead to colossal leaps in scientific progress.
 
Tragically, we’ve been in a period of acute extremity for some time now. And though it may be of scant consolation, there have been advances across all sorts of industries as a result. The one that’s closest to home for Teradata, however, is the application of data science to healthcare.

A period of unique breakthroughs  

Earlier this year, Teradata staff across the world congratulated our teammate Carlos Ortega for developing an algorithm to help battle Covid-19. Carlos is the Principal Data Scientist in our Madrid office, and spent most of his Christmas holiday working on the AI model. It serves as a diagnostic tool to help doctors in their decision making when treating coronavirus patients. And it will soon be passed over to the Covid-19 Host Genetics Initiative, which is a project aimed at bringing together the world's genetic community to share and analyse data on the virus.
 
In the same way that everybody knew apples fall from trees before Isaac Newton observed it – applying data science to healthcare is not a new concept. Even before the pandemic, there was talk of using AI and machine learning to collect, structure and process high volumes of medical data, and use it to extract insights that could save lives.
 
The point is that circumstances conspired to place Newton in his parents’ orchard at precisely the moment he had the time and space to form his theory of gravity. Likewise, Covid-19 and successive lockdowns may be the unique context in which gigantic advances in data science and medicine can take place.
 
The achievement of our colleague, Carlos, is just one story. There are many other cases where the urgent Covid-19 situation has pushed data scientists to innovate. Take the researchers at the Max Planck Institute for example. They have developed an algorithm to predict the individual mortality risk for patients with Covid-19. It uses machine learning methods to analyse data from thousands of patients around the world. And it’s named COVEWS, which stands for ‘Covid-19 Early Warning System’. Now the researchers are confident they can re-train the algorithm to predict mortality risks for other diseases. Which means it will continue to be a lifesaver, even after Covid-19 has been defeated. 
 
There have been lots of similar efforts to develop AI and machine learning tools to help doctors with Covid-19 diagnoses and patient triage. And supranational bodies like the European Union and the World Health Organisation have been pivotal.
 
For a wholly different breakthrough, however, we don’t need to leave the Max Planck Institute. Last year, they also developed a model to calculate the coronavirus infection risk from aerosol transmission in indoor spaces. It uses variables like: the size of the room, the number of people in it and what they’re doing – to estimate both the chances of Covid-19 infection and the risk to anyone in that room. It also estimates how protective measures like masks and ventilation will reduce the risk.
 
Elsewhere, a team at the Massachusetts Institute of Technology (MIT) has developed an algorithm which can identify people with Covid-19, simply by listening to the sound of their coughs. The crucial differences between a Covid cough and non-Covid cough are inaudible to human ears. But by analysing around 70,000 audio samples, the MIT AI model has learned to pick up the infinitesimal variations. And with a staggering 98.5% success rate, too.
 
Another team at MIT has created a machine-learning approach to identifying existing drugs which can be repurposed to fight coronavirus in elderly patients. And on the other side of the USA, in Santa Clara, a 27 year-old data scientist named Youyang Gu developed his own Covid-19 forecasting model. And even though he created it in his bedroom, the model made more accurate predictions than those coming from grand institutions with even grander budgets.

Is this just the beginning?

Just like the Teradata Covid-19 Resiliency Model, all these stories exhibit the enormous potential to do good with data. And most – if not all – share a common trait. Not only did they come about because of the Covid-19 crisis, but they open up all sorts of possibilities to make the world a better place, even once the crisis is over.
 
The end of the pandemic may well be in sight, but it’s highlighted the incredible power of data science to transform economies, industries and people’s lives for the better. Just like Isaac Newton in lockdown at his parents’ manor, we may have witnessed the start of something much, much greater.